March 26, 2001, 4:21 PM —
Why shadow passwords? Simply put, the shadow password scheme addresses the major shortcoming of the original Unix password-handling scheme, the fact that the password list was stored as a world-readable file.
The encoding mechanism for Unix passwords was (and is) very secure, being a one-way algorithm and therefore easy to apply but impossible to reverse. However, the password file itself is vulnerable to a cracking technique known as a dictionary attack, in which all the words from a large dictionary file are encoded and compared with the encoded password (readable by any user, remember) in /etc/passwd. This dictionary file is usually based on a normal English-language dictionary, with the addition of slang and weak passwords like "gandalf," "xyzzy," "qwerty," or even (God help us) "password." If the two match, then the original unencoded word is the password.
This may sound simple, but it takes a while to run the tens, or hundreds, of thousands of dictionary entries against a single password. Still, it is not extremely difficult with today's high-performance computing systems. Shadow passwords retain the Unix password mechanism and its backward compatibility with the huge Unix application base, while preventing the dictionary attack.
